Specify violation messages in input files

[Vyom-Yadav]

Single Check might have multiple violations messages, we need to see exact violation message if Check has multiple messages, to be sure that violation of correct type is placed at right line.

Violation messages have not been specified in certain input files, a list of checks in which violation message is not specified in the input file and have more than one violation message key has been specified here-

checkstyle/src/test/java/com/puppycrawl/tools/checkstyle/bdd/InlineConfigParser.java

Line 99 in 72f1822

private static final Set<String> SUPPRESSED_CHECKS = new HashSet<>(Arrays.asList(

To Do-

• Pick a check (single Check Inputs files update in single Pull Request) To allocate Check update for you please make comment in issue "I am on XxxxxCheck"
• Find all input files that use that check
• Add violation message at the end of the // violation comment
• Remove the check from the suppression list

Same has been demonstrated at 461cd04
If the violation message contains special regex characters, escape that character.
Ignore input files using more than a single check.

Ignore lines with multiple violations, eg-

// 2 violations
// 2 violations above
// 4 violations below

and any other violation message of this nature.

List of Checks:

• RegexpCheck
• EmptyForInitializerPadCheck
• JavadocStyleCheck
• SummaryJavadocCheck

Old tracker (some might be outdated):

• RightCurlyCheck
• EqualsAvoidNullCheck
• MethodParamPadCheck
• DescendantTokenCheck
• RequireThisCheck
• EqualsHashCodeCheck
• JavaNCSSCheck
• ReturnCountCheck
• AnnotationUseStyleCheck
• EmptyLineSeparatorCheck
• OrderedPropertiesCheck
• AbstractJavadocCheck
• DeclarationOrderCheck
• MissingOverrideCheck
• AbstractClassNameCheck
• MethodCountCheck
• PackageDeclarationCheck
• JavadocTypeCheck
• LeftCurlyCheck
• RegexpCheck
• WhitespaceAfterCheck
• MultipleVariableDeclarationsCheck
• EmptyForInitializerPadCheck
• UnnecessaryParenthesesCheck
• JavadocMethodCheck
• ModifierOrderCheck
• MissingDeprecatedCheck
• RegexpOnFilenameCheck
• RedundantImportCheck
• EmptyBlockCheck
• UniquePropertiesCheck
• WhitespaceAroundCheck
• OperatorWrapCheck
• GenericWhitespaceCheck
• EmptyForIteratorPadCheck
• JavadocStyleCheck
• JavadocPackageCheck
• FallThroughCheck
• WriteTagCheck
• TranslationCheck
• FileTabCharacterCheck
• ImportOrderCheck
• JavadocContentLocationCheck
• JavadocParagraphCheck
• SummaryJavadocCheck
• DefaultComesLastCheck
• AnnotationLocationCheck
• CustomImportOrderCheck
• NewlineAtEndOfFileCheck
• VariableDeclarationUsageDistanceCheck
• ImportControlCheck

format it like:

checkstyle/src/test/resources/com/puppycrawl/tools/checkstyle/checks/coding/declarationorder/InputDeclarationOrderOnlyModifiers.java

Line 16 in 5b2a0ef

public static final int FOO = 3; // violation 'Variable access definition in wrong order.'

if message is long, we can shorten it by regexp usage (but do not over use regexp, message should be readable):

checkstyle/src/test/resources-noncompilable/com/puppycrawl/tools/checkstyle/checks/coding/equalsavoidnull/InputEqualsAvoidNullEnhancedInstanceof.java

Line 15 in 4bd9856

boolean myBool = myString.equals("MyString!!"); // violation 'left .* of .* equals'

There are a lot of examples in code just search for // violation in Input files for more examples.

[Kevin222004]

I am on WhitespaceAfterCheck

[SarveshLimaye]

I am on EmptyBlockCheck

[Kevin222004]

I am on WhitespaceAfterCheck
@Vyom-Yadav I have found 13 input files contain the message // violation. can I sharing a image to know that i am on right path or not

[Vyom-Yadav]

@Kevin222004 Sure

[Vyom-Yadav]

@Kevin222004 currently we have // violation comment, you have to replace it with the actual violation message, you can see the violation message from respected test file, you can use 461cd04 as reference.

[SarveshLimaye]

@Vyom-Yadav I am not able to find actual violation msg. Can you pls help me with it ?

[Kevin222004]

@Kevin222004 currently we have // violation comment, you have to replace it with the actual violation message, you can see the violation message from respected test file, you can use 461cd04 as reference.

@Vyom-Yadav can I go through this documentation
https://checkstyle.sourceforge.io/config_whitespace.html#WhitespaceAfter

[Vyom-Yadav]

@SarveshLimaye It will be in the .properties file, if you are not able to find it, just search within the whole project (Ctrl + Shift + F in intellij).

@Vyom-Yadav can I go through this documentation

Yes, you can read whatever you wish to.

[SarveshLimaye]

@Vyom-Yadav Thanks a lot! Can you confirm that this is the right .properties file
https://github.com/SarveshLimaye/checkstyle/blob/master/src/main/resources/com/puppycrawl/tools/checkstyle/checks/blocks/messages.properties

[Vyom-Yadav]

Kindly make PR's so we can see better, every check has a violation message MSG_{someMsg} which is equal to something like block.empty or some other short key that finally points to the complete message.

MSG_KEY -> msg.key -> This is the message key.
<-------------------> 
          <------------------------------------>
Check                  .properties file

The overlapping portion is present in both the check and the .properties file.
Ctrl + click will take you to the declaration in intellij.

[Kevin222004]

@Vyom-Yadav How to add violation message which contain more than one violation (ex :- // 2 violation and // 3 violation )

[Vyom-Yadav]

@Kevin222004 You don't have to specify a violation message where there are multiple violations, you can ignore those cases, I updated PR desc to demonstrate the same.

[Kevin222004]

@Vyom-Yadav

i am finding difficulties to solve this error

[Vyom-Yadav]

The line shouldn't be longer than 100 characters. If your violation message is crossing 100 characters, eg-

int a = 12; // violation, some loooooooooooooooooong violation message

you can change it to-

int a = 12;
// violation above, some looooooooooooooong violation message

to reduce the line length.

[romani]

better to keep traling comment on same line as violation
int a = 12; // violation, some lo*ng violation message

as message supports regexp and allow us to remove some obvious parts.

[Vyom-Yadav]

@Kevin222004 It's your decision, you have to decide what to do based on some factors, first ask yourself if I shorten this message with the help of regex, will it make sense? secondly if not then try wrapping the violation part on the next line to fall within the limits, there is no set of rules, you have to decide what to do, in this particular case I would have wrapped the second argument on the next line to keep the violation message clearer.

[romani]

@Kevin222004 , you can shorten names parameters, change used types of parameter, you can do line wrap(but it will force you do more updates in test expectation due to line number shift.).
We might do removal of first argument, as it is not required for this validations.

[SarveshLimaye]

@Vyom-Yadav @romani I am not able to solve this errors

[romani]

please send PR to let us see your problem and exact code (all details matter) that you have.

[SarveshLimaye]

@romani have already raised pr. Could you pls take a look . #11322

[faraazb]

Hi! I am on UnnecessaryParenthesesCheck. I used the corresponding test file to identify the input files. Will make a PR soon!

[suniti0804]

In the class src/test/java/com/puppycrawl/tools/checkstyle/checks/javadoc/JavadocStyleCheckTest.java it's unusual to find keyword "violation" in the expected array. For other checks that we have, we don't have this keyword in the expected array itself and only in the input files. Can I please get an understanding on this?

An example -

@Test
   public void testExcludeScope1()
           throws Exception {
       final String[] expected = {
           "23: " + getCheckMessage(MSG_NO_PERIOD),
           "48: " + getCheckMessage(MSG_NO_PERIOD),
           "56:11: " + getCheckMessage(MSG_UNCLOSED_HTML,
                   "<b>This guy is missing end of bold tag // violation"),
           "59:7: " + getCheckMessage(MSG_EXTRA_HTML, "</td>Extra tag "
                   + "shouldn't be here // violation"),
           "60:49: " + getCheckMessage(MSG_EXTRA_HTML, "</style> // violation"),
           "61:19: " + getCheckMessage(MSG_UNCLOSED_HTML, "<code>dummy. // violation"),
           "70: " + getCheckMessage(MSG_NO_PERIOD),
           "71:31: " + getCheckMessage(MSG_UNCLOSED_HTML, "<b>should fail // violation"),
           "92:39: " + getCheckMessage(MSG_EXTRA_HTML, "</img> // violation"),
       };
       verifyWithInlineConfigParser(
               getPath("InputJavadocStyleExcludeScope1.java"), expected);
   }

The expected array comes out to be as -

23: First sentence should end with a period.
48: First sentence should end with a period.
56:11: Unclosed HTML tag found: <b>This guy is missing end of bold tag // violation
59:7: Extra HTML tag found: </td>Extra tag shouldn't be here // violation
60:49: Extra HTML tag found: </style> // violation
61:19: Unclosed HTML tag found: <code>dummy. // violation
70: First sentence should end with a period.
71:31: Unclosed HTML tag found: <b>should fail // violation
92:39: Extra HTML tag found: </img> // violation

[romani]

@suniti0804 ,

it's unusual to find keyword "violation"

Thanks a lot for raising this. It is bad design of our old tests.
We use to have this before we added support of // violation X lines below

Ideally we should refactor such tests to remove violation comment from javadoc and put them above javadoc.

[suniti0804]

@suniti0804 ,

it's unusual to find keyword "violation"

Thanks a lot for raising this. It is bad design of our old tests. We use to have this before we added support of // violation X lines below

Ideally we should refactor such tests to remove violation comment from javadoc and put them above javadoc.

To address this, shall we make a new issue or we can make the fix here only in the issue #11214 ?

[romani]

Please create new issue to update all javadoc Checks to avoid // violation in javadoc comments

[suniti0804]

src/test/java/com/puppycrawl/tools/checkstyle/checks/javadoc/JavadocStyleCheckTest.java

We have this issue now - #13969

[suniti0804]

Some of the input files using the class - FinalClassCheckTest doesn't have violation message specified in them and yet this check is not in the suppression list and evidently not causing any errors as well. What could be the reason for this?

[romani]

Not all Input* files has violations, as some files are absolutely show how code should looks like to have no violations in it.
All methods in this Test file use verifyWithInlineConfigParser that is the best form of test we should have. Some tests without such method covers extreme edge cases.

[sushant1987]

I am on JavadocStyleCheck

[Vyom-Yadav]

Thanks a lot to all the contributors involved!